The seduction of building AI tools to make work easier
And the line between building something useful and building for building's sake
A few weeks ago, I vibe-coded a new website for myself. If you’re not familiar with vibe coding, it’s using natural language (“Hey Claude, we’re going to work on a new homepage for me”) to code something. An article in Microsoft Source describes it this way:
Coding has long been limited to the realm of software engineers who studied it in school. But now there are so-called no-code, low-code and pro-code AI tools. That’s broadening who can build apps, helping everyone from non-technical business professionals to experienced developers solve problems, save time and boost creativity.
It’s addicting… after building my initial website, I spun up several more. I’ve spent hours fiddling with Claude CoWork. It’s exciting and genuinely a big deal for people who like to “tinker.” You can now build the tool you’ve always wanted — either because no company was going to make it (because the use case is too specific to you), or because you don’t want to pay a monthly subscription for something you’ll only partially use.
All of this has a real impact on people’s lives, both at work and personally. But the ability to build doesn’t mean you should build everything. And the line between building something useful and building for building’s sake is thinner than most of us realize.
The new math of building
Companies have always wrestled with the “build vs. buy” decision. If they wanted a tool to do something, they could either build it themselves (which takes time, money, and resources) or buy it (which lacks full control and customization). Now the calculation is murkier, and it can rely on individual decisions instead of a big corporate buy-in. You want something? You can build it yourself. The question is whether you should.
What used to require hiring a designer, a developer, or a consultant can now be done in a few hours with an AI tool. That’s a real shift, and for many people, it’s the first time they’ve been able to bring an idea to life without depending on someone else.
Regular people are building personal websites, scheduling tools, custom dashboards, automations that connect their apps — things that weren’t possible a year ago. The potential is enormous.
But the potential also has a way of expanding to fill all available time, rather than saving time.
Katie Parrott wrote a piece called “AI Was Supposed to Free My Time. It Consumed It.” She describes staying up until 1 a.m., building and rebuilding with her AI assistant. Her piece captures something important: a compulsive quality of AI-assisted building that many people find once they get started. It doesn’t feel like work. It feels like momentum… and that’s what makes it hard to stop. (You can read Katie’s guest essay about productivity for this publication here.)
There’s a version of this playing out on a much larger scale. On a recent episode of The Verge’s Decoder podcast, Hasbro CEO Chris Cocks described how AI-assisted design has transformed the company’s product development. Where Hasbro’s teams used to take one or two toy concepts to the full prototype phase, they can now take 10 or 20. More ideas, faster, at the same cost.
That sounds like pure upside. But it’s also a perfect illustration of Parkinson’s law, a concept based on a satirical essay that contained the line, “Work expands so as to fill the time available for its completion.” When AI makes it possible to generate 20 prototypes instead of two, you don’t stop at two and go home early. You generate all 20. And then you review all 20. And then you iterate on the five most promising ones. The capacity to do more becomes the expectation to do more.
This pattern shows up in survey data, too. The Upwork Research Institute surveyed 2,500 workers in 2024, including executives, employees, and freelancers. 77% of employees said AI tools had actually added to their workload. Workers reported spending more time reviewing AI-generated output, more time learning the tools, and being asked to do more work as a direct result of the technology. More recently, Upwork’s 2025 follow-up study found that of the workers reporting the highest productivity gains from AI, 88% of them experienced burnout.
This cycle of constant building with AI only benefits workers if they were previously overworked, and AI actually reduces the load. If the “newfound time” just gets filled with more stuff (more prototypes, more iterations, more output), then the gains flow to the employer. The worker doesn’t get any time back.
Building with intention
This isn’t an argument against AI tools. I’m a fan, trust me. But it’s an observation that the low friction to get started doesn’t mean you should build everything. It’s no longer a “build vs. buy” question, but instead a “build vs. ‘what do I gain from building?’” question.
The temptation to build everything you can build is real. And the people who benefit most from these tools are the ones who build the right things: the things that actually improve their lives rather than just consuming their attention.
I could stay up allll night (like my friend Katie) and build new things in CoWork or Claude Code. Instead, I’ve got to decide how to meaningfully apply my time and attention. If I were working for an employer, it would be the same.
Before you sit down to build something with AI (whether it’s a custom app, an automation, or a new workflow), a few questions are worth asking:
What specific problem does this solve? If you can’t name a real use case — something that’s actually costing you time, money, or frustration on a regular basis — it might be a distraction posing as a productivity gain.
What’s the real cost? Not in dollars, but in hours. How much time will you spend getting this up and running versus how much time it will actually save you? A tool that takes 10 hours to build and saves you 20 minutes a week will take a while to “pay for itself” (though that certainly compounds if you save 20 minutes a week forever).
When will I stop? Deciding on an end result before you start is fundamentally different from trying to stop once you’re in the flow. (The flow is a trap!)
What will I do with the time I get back? This is the question that keeps the whole building process in check. If AI takes something off your plate, the benefit only exists if you’re intentional about what fills that space. Otherwise, you’ve just replaced one type of busy with another.
Honestly, I’m haven’t reached the point where building gives me more time… yet. I’m too deep into changing processes. Some are things I’ve had in place for years and are forcing me to fundamentally change how I do work. Will it pay off, eventually? That’s the bet I’m making: that this upfront work will be worth something.
Want to build a life-first business? These reflections will help you determine your priorities.
If you want to support my work as a writer, you can subscribe to receive additional issues I publish.
Have a work story you’d like to share? Please reach out using this form. I can retell your story while protecting your identity, share a guest post, or conduct an interview.
Good article. I think we should really focus on the timesavings. That would be beneficial for all of us. Gender equality, health etc.